Organised armed force between opposing parties.
ANALYSIS/INTERNAL CONFLICT
SESSION 2:1UR · EESPI
Analysis of Internal Conflict
School of Social, Political and International Studies
Universidad del Rosario
Mathew H. Charles, PhD
Session 2:1 · What Is Civil War?
What is
civil war?
Definitions, boundaries and the politics of classification
02 THE BASIC QUESTION
When does internal violence become civil war?
We need criteria—not simply an impression that the violence is severe.
A broader category that includes repression, riots, terrorism and war.
Organised coercion linked primarily to illegal markets and protection.
THE CLASSIFICATION PROBLEMWhat combination of organisation, political purpose, armed combat and intensity distinguishes civil war?
By the end: you should be able to classify a case, defend your reasoning and explain what evidence would change your answer.
03 ONE REALITY, DIFFERENT FRAMEWORKS
There is no single label used by everyone
The terms overlap, but they are not interchangeable: different institutions use them to answer different questions.
CIVIL WAR
An analytical and public category
- Used by
- Political scientists, historians, journalists and political actors
- Main question
- Has domestic political conflict become organised war over sovereign authority?
- Key caution
- There is a shared conceptual core, but no universally accepted measurement rule.
INTERNAL / INTRASTATE ARMED CONFLICT
A policy and research category
- Used by
- Governments, policy institutions and conflict-data projects
- Main question
- Is organised armed conflict occurring mainly within a state?
- Key caution
- The precise threshold depends on the institution or dataset using the term.
NON-INTERNATIONAL ARMED CONFLICT · NIAC
A legal classification
- Used by
- The ICRC, courts, humanitarian organisations and military legal advisers
- Main question
- Do the organisation and intensity of hostilities trigger international humanitarian law?
- Key caution
- Political motivation is not required, and there is no fixed fatality number.
Not necessarily. A lawyer may classify a NIAC to determine which humanitarian rules apply, while a political scientist asks whether the same violence belongs in a civil-war dataset. A government may choose “internal armed conflict” for policy or political reasons. Always ask who is using the term, for what purpose and according to which test.
Source: Kalyvas 2007, pp. 416–418; ICRC 2024; UCDP Definitions 2026.
04 FIRST IMPRESSION
Which of these situations would you call a civil war?
Read the eight country situations first. Then choose—before we introduce a formal definition.
SudanOrganised forces fight for control of the state amid foreign involvement and several armed fronts.
SyriaA long civil war changes after the fall of a government, while several armed dyads remain active.
ColombiaA peace agreement ends one central dyad, but other insurgent, dissident and criminal actors continue fighting.
MexicoHighly organised groups use military capabilities, govern territory and generate extreme levels of violence.
HaitiArmed coalitions control neighbourhoods and challenge state capacity, while their political objectives remain contested.
MyanmarThe military government fights the National Unity Government’s forces and several long-established ethnic armed organisations across multiple fronts.
YemenDomestic forces contest government and territory while foreign states intervene and additional interstate confrontations overlap with the internal conflict.
DR CongoGovernment forces, organised domestic armed groups and externally supported actors fight across several territories and armed relationships.
Discuss first
Can one label fit all eight cases?
Source: Case descriptions informed by UCDP data for 2025; classifications are teaching interpretations.
05 WHY WORDS MATTER
“Civil war” is never just a neutral description
The label can alter legitimacy, law, policy and the imaginable response.
Legitimacy
Does an armed organisation appear to be a political party to a conflict—or merely a collection of offenders?
Response
Does the situation call for policing, military action, humanitarian protection, negotiation—or several at once?
Evidence
Which cases enter a dataset, and which disappear because they fall outside the coding rules?
Memory
How will later generations name the violence and assign responsibility?
A government may prefer labels such as emergency, terrorism, banditry or organised crime because “civil war” can appear to recognise an opponent’s political and military standing.
Source: Armitage 2017, Introduction; Kalyvas 2007, pp. 416–417.
06 THREE DIFFERENT JOBS
A definition can name, explain or measure
These tasks overlap, but they are not the same.
POLITICAL LABEL
What shall we call it?
A contested public description used by governments, armed actors, journalists and citizens.CONCEPTUAL DEFINITION
What makes it this kind of phenomenon?
An analytical account of the essential features that distinguish civil war from related violence.OPERATIONAL DEFINITION
How will we decide case by case?
Observable rules that allow researchers to include or exclude cases consistently.“Poverty” is a concept. A particular income line or multidimensional index is an operational rule for measuring it. In the same way, “civil war” is a concept, while a fatality threshold is one possible coding rule.
07 ARMITAGE
Civil war has a history—and naming is political
The concept developed through centuries of argument rather than arriving as a timeless scientific category.
ROMEbellum civile
War among citizens created a conceptual problem: enemies were simultaneously insiders.
LATER USESRevolution, rebellion, sedition
Competing words framed similar violence differently and assigned different legitimacy.
TODAYA global category
The term travels across contexts, but its meaning and consequences remain disputed.
Armitage’s lesson: before asking whether a conflict “is” a civil war, ask who is naming it, from which position and with what consequences.
Source: Armitage 2017, Introduction.
08 DISCUSSION LAB
The same violence looks different from different positions
Select a perspective, then compare its incentives and blind spots.
From this position
Government
Calling opponents ‘criminals’ or ‘terrorists’ can deny them political standing and suggest that policing—not negotiation—is appropriate.
2 MINUTES
How does each actor’s choice of label change how the conflict is understood—and which response appears appropriate?
Source: Armitage 2017, Introduction; Kalyvas 2007, p. 416.
09 KALYVAS
Begin with organised domestic armed combat
Kalyvas offers a conceptual definition before turning to measurement.
Civil war is
organised armed combatinside a recognised sovereign entity
between parties under a common authoritywhen hostilities began.
Why “common authority”?
It distinguishes a domestic challenge from a war that begins between separate sovereign states.
Source: Kalyvas 2007, pp. 417–418.
10 BUILD THE CONCEPT
Five questions unpack the definition
A difficult case should be tested element by element—not classified by instinct.
Armed combat
Is organised force being used between opposing parties?
Competing sides
Is there a substantial armed organisation, not only isolated attackers?
Domestic relationship
Were the parties subject to a common authority at the beginning?
Sovereign challenge
Is authority over government or territory being challenged?
Sustained organisation
Can the opposition recruit, command and mount continuing resistance?
Source: Kalyvas 2007, pp. 417–418.
11 THE BOUNDARIES
Civil war is related to—but not identical with—other violence
These are analytical distinctions. Real conflicts can combine several forms.
Riot
Collective violence that may be intense but lacks sustained military organisation.
Terrorism
A tactic or form of violence; an organisation using it may or may not also participate in civil war.
Civil war
Organised domestic armed combat involving a challenge to sovereign authority.
Organised crime
Organised coercion for illegal markets or protection, without necessarily contesting sovereign authority.
One-sided violence
Organised attacks against civilians rather than reciprocal combat between warring parties.
No. “Terrorism” normally identifies tactics or targets; “civil war” classifies a wider relationship between organised parties. A group can use terrorism within a civil war.
Source: Kalyvas 2007, pp. 417–418; UCDP Definitions 2026.
12 SAMBANIS
An operational definition turns the concept into coding rules
Researchers must decide what evidence is sufficient to place a case in a dataset.
CONCEPTCivil war
→OBSERVABLE EVIDENCEActors · combat · deaths · duration
→CODING DECISIONInclude · exclude · flag as ambiguous
01 Does the state participate as a principal combatant?
02 Are the opposition and its political objectives organised?
03 Can local rebels mount effective resistance?
04 Does the violence cross an intensity threshold?
05 When does one episode begin, end or restart?
Source: Sambanis 2004; see his proposed coding rules.
13 THRESHOLD LAB
A death threshold is a rule—not a natural law
Move the threshold and watch the dataset change.
500battle-related deaths
255001,000
Case A40 deathsEXCLUDED
Case B180 deathsEXCLUDED
Case C420 deathsEXCLUDED
Case D760 deathsINCLUDED
Case E1,200 deathsINCLUDED
With the rule set at 500 deaths, 2 of 5 fictional conflicts enter the dataset.
Case A, Case B, Case C are excluded because they fall below the selected minimum.
Why did the result change? Moving the slider changes the researcher’s rule, not the violence. Moving right makes the rule stricter, so more lower-fatality conflicts disappear from the dataset.
Illustrative cases only. These numbers do not represent real conflicts.
UCDP “armed conflict”At least 25 battle-related deaths in a dyad during a calendar year
UCDP “war” intensityAt least 1,000 battle-related deaths in a calendar year
Sambanis’s problemDifferent thresholds and counting rules produce different lists of civil wars
Source: Sambanis 2004; UCDP Definitions 2026.
14 SAME REALITY, DIFFERENT DATA
Coding choices can change the answer
That does not make data useless. It makes definitions visible and testable.
Research choiceOption AOption BPossible consequence
DeathsBattle deaths onlyInclude civilian deathsDifferent intensity
TimeAnnual thresholdCumulative thresholdDifferent onset
ResistanceAny organised fightingState must suffer lossesDifferent cases
PauseOne intermittent warSeveral episodesDifferent duration
Because explicit rules allow comparison, replication and criticism. The correct response is not to abandon data, but to test whether conclusions survive reasonable alternative definitions.
Source: Sambanis 2004.
15 TIME
When does one civil war begin—and when does it end?
Peace agreements, pauses and splinter groups create coding problems.
1Escalation
Organised armed confrontation begins.
2Threshold crossed
The case enters a particular dataset.
3Ceasefire
Combat falls—but the incompatibility may remain.
4Splinter violence
A faction continues or forms a new dyad.
5What ended?
The war, one episode, or only one relationship?
Discuss first
After a peace agreement, is later violence automatically the same civil war?
Source: Sambanis 2004; UCDP Definitions 2026.
16 A NEW QUESTION
Have civil wars themselves changed?
The post–Cold War period produced a powerful claim: contemporary wars form a distinct type.
BEFORE 1989?
Ideological armies
Centralised organisations, political programmes, disciplined forces and clearer fronts.
VS
AFTER 1989?
Networked violence
Militias, identity politics, illicit finance, attacks on civilians and blurred domestic–international boundaries.
Discuss first
Is this a real historical transformation—or an oversimplified contrast?
Source: Kaldor 1999; Kalyvas 2001, pp. 99–102.
17 KALDOR
The “new wars” thesis identifies changed goals, methods and finance
It links contemporary organised violence to globalisation, fragmented authority and transnational networks.
Identity politics
Mobilisation around exclusive ethnic, religious or communal identities rather than universal projects.
Networks
State forces, militias, paramilitaries, mercenaries, criminal organisations and external supporters overlap.
Control through fear
Displacement and violence against civilians become central methods of political control.
Predatory economies
Looting, extortion, illicit trade, diaspora contributions and foreign assistance can sustain violence.
Important: this is a theoretical thesis to examine—not a description that can be applied automatically to every post-1989 conflict.
Source: Kaldor 1999.
18 COMPARE THE CLAIM
“Old” and “new” wars are usually contrasted along three dimensions
Kalyvas reconstructs the argument before testing whether the contrast survives historical evidence.
DimensionSupposedly “old”Supposedly “new”
Motivation
Collective grievance and ideology
Private gain, predation and loot
Support
Broad popular support
Little support; dependence on coercion
Violence
Controlled and disciplined
Gratuitous and undisciplined
The contrast idealises earlier wars and interprets contemporary wars using incomplete, highly visible information. A neat table can organise a claim without proving that the categories accurately describe history.
Source: Kalyvas 2001, pp. 101–102.
19 KALYVAS’S CRITIQUE
Contemporary wars may differ—but not through a clean historical break
Kalyvas tests the thesis against research on both recent and earlier civil wars.
Old wars were not purely ideological
Local rivalries, coercion, private motives and material interests also shaped participation in earlier conflicts.
New wars are not necessarily depoliticised
Illicit finance and fragmented organisation do not erase political projects or struggles over authority.
Extreme violence is not new
Militias, massacres, forced displacement and child recruitment also appear in earlier civil wars.
Conclusion: investigate variation directly. Do not infer the nature of a war simply from whether it occurred before or after 1989.
Source: Kalyvas 2001, pp. 99–118; Kalyvas 2007, p. 426 n. 6.
20 EVIDENCE TEST
Separate three questions that are often collapsed
Funding, individual motives and organisational objectives are related—but not identical.
How is violence financed?
Taxation, foreign support, extortion, coca, mining, kidnapping or donations.
Why do individuals participate?
Belief, coercion, revenge, protection, belonging, income or local rivalry.
What authority does the organisation pursue?
Government, secession, autonomy, local rule, protection of markets—or a combination.
An organisation can finance a political project through illegal markets. Conversely, political language does not prove that every member—or commander—is motivated by ideology.
Source: Kalyvas 2001, pp. 102–108; Kalyvas 2007, pp. 420–421.
21 POLITICAL OR CRIMINAL?
The categories can overlap without becoming meaningless
The task is to identify the relationship between economic extraction and political authority.
POLITICALChange government
Control territory
Regulate populations
Control territory
Regulate populations
CRIMINALControl markets
Extract profit
Evade law enforcement
Extract profit
Evade law enforcement
OVERLAPArmed governance
Extortion
Territorial control
Extortion
Territorial control
Discuss first
Does drug trafficking make an armed organisation non-political?
Source: Kalyvas 2001, pp. 102–108; Sambanis 2004.
22 COLOMBIA AS A TEST
Coca dependence did not automatically make the FARC a criminal enterprise
Kalyvas uses Gutiérrez’s Colombian research to challenge the “criminal rebels” shortcut.
“Coca finances the group, therefore the group is criminal rather than political.”
- How were resources controlled and distributed?
- What conduct did the organisation permit or punish?
- What political authority did it seek to exercise?
- Why did members join, remain or leave?
Teaching point: a plausible label is not a demonstrated causal mechanism. Case research can reveal what aggregate categories hide.
Source: Gutiérrez Sanín 2004; discussed in Kalyvas 2007, p. 420.
23 A BETTER TYPOLOGY
Kalyvas distinguishes forms of warfare and forms of state challenge
These dimensions can be more useful than dividing wars into “old” and “new.”
Army confronts army
Relatively symmetric forces hold positions and engage more directly.
State versus insurgency
A militarily weaker organisation relies on guerrilla tactics, concealment and civilian networks.
Militia confronts militia
Fragmented forces and loose command structures accompany state implosion.
Peripheral challengeAuthority is contested from areas of limited central control.
State implosionCentral institutions and armed forces fragment into competing centres.
Source: Kalyvas 2007, pp. 427–430.
24 LAYERED CONFLICT
One civil war can contain many conflicts
A national “master cleavage” may conceal local, economic and social struggles.
NATIONAL POWERGovernment ↔ armed opposition
RELIGION & IDENTITYMultiple, changing collective boundaries
LOCAL RIVALRIESCommunity, clan and personal conflicts
ECONOMY & RESOURCESLand, extraction, routes and inequality
EXTERNAL CONNECTIONSNeighbouring states, diasporas and sponsors
That question may be misleading. Different layers can interact and change one another. Good analysis explains how national and local dynamics combine instead of reducing the conflict to one master cause.
Source: Kalyvas 2007, pp. 425–427, using Sudan as the layered example.
25 BOUNDARY-CASE LAB
Classify, justify, then identify what is missing
A defensible answer may be “uncertain”—but only after explaining why.
Strong civil-war case
Sudan
Organised forces fight for control of the state amid foreign involvement and several armed fronts.
Source: Case framing informed by UCDP 2025 material; classifications are teaching interpretations.
26 INTERACTIVE MAP
Organised violence in 2025: which cases count as civil war?
The map offers selected cases for classification—not a pre-labelled list of civil wars.
Teaching map, not an exhaustive list. Cases are selected from the UCDP 2025 conflict material to practise classification; UCDP records dyads and types of organised violence, not one unquestionable list of “civil wars.”
Source: UCDP/PRIO Armed Conflict Dataset v26.1; Davies, Pettersson and Öberg 2026; UCDP Definitions 2026.
27 DECISION GUIDE
Use the same sequence for every case
The result is a reasoned classification, not a guess.
Organised armed combat?
If no: examine protest, riot, isolated terrorism or criminal violence.
Domestic relationship at the outset?
If no: examine interstate or extra-systemic conflict.
Challenge concerning government or territory?
If unclear: investigate stated and practised objectives.
State and organised opposition as combatants?
If neither side is the state: examine non-state conflict.
Required intensity and duration?
State the dataset or threshold you are using.
One war, several dyads or several episodes?
Disaggregate actors, issues, territories and years.
Source: Synthesis of Kalyvas 2007; Sambanis 2004; UCDP Definitions 2026.
28 FINAL SYNTHESIS
What we can—and cannot—conclude
A careful definition strengthens an argument by limiting it.
- Identify the criteria used to classify a conflict.
- Compare cases using consistent rules.
- Explain why a case sits near a boundary.
- Test whether results change under another reasonable definition.
- Recognise political and criminal dynamics operating together.
- Treat a death threshold as a natural dividing line.
- Assume every violent internal crisis is a civil war.
- Infer political purpose from intensity alone.
- Infer criminal purpose from illicit finance alone.
- Assume post–Cold War conflicts are automatically “new wars.”
A civil war is more than severe violence inside a country: it is sustained, organised domestic armed combat involving a challenge to sovereign authority. Its boundaries still require explicit evidence and judgement.
Source: Synthesis of the session readings.
29 VERIFIED READING LIST
References used in this presentation
The bibliography distinguishes the assigned readings, the new-wars debate and the source for current conflict classifications.
- Armitage, David. 2017. Civil Wars: A History in Ideas. New York: Alfred A. Knopf. Introduction.
- Davies, Shawn, Therése Pettersson and Magnus Öberg. 2026. ‘Organized violence 1989–2025, and violent political protests.’ Journal of Peace Research. https://doi.org/10.1093/jopres/xjag046.
- Gutiérrez Sanín, Francisco. 2004. ‘Criminal Rebels? A Discussion of Civil War and Criminality from the Colombian Experience.’ Politics & Society 32(2): 257–285.
- International Committee of the Red Cross (ICRC). 2024. How Is the Term ‘Armed Conflict’ Defined in International Humanitarian Law? Opinion Paper, April 2024.
- Kaldor, Mary. 1999. New and Old Wars: Organized Violence in a Global Era. Stanford, CA: Stanford University Press.
- Kalyvas, Stathis N. 2001. ‘“New” and “Old” Civil Wars: A Valid Distinction?’ World Politics 54(1): 99–118. https://doi.org/10.1353/wp.2001.0022.
- Kalyvas, Stathis N. 2007. ‘Civil Wars.’ In Carles Boix and Susan C. Stokes, eds., The Oxford Handbook of Comparative Politics, 416–434. Oxford: Oxford University Press.
- Sambanis, Nicholas. 2004. ‘What Is Civil War? Conceptual and Empirical Complexities of an Operational Definition.’ Journal of Conflict Resolution 48(6): 814–858. https://doi.org/10.1177/0022002704269355.
- Uppsala Conflict Data Program (UCDP). 2026. UCDP Definitions and UCDP/PRIO Armed Conflict Dataset, version 26.1. Department of Peace and Conflict Research, Uppsala University.
Referencing rule used here
Author–year–page references appear on teaching screens when a claim comes from a specific passage. Current classifications are presented cautiously and tied to UCDP’s 2025 data and definitions.
01 / 29